Revisit parents/children of terms to make sure none were missed

When using a limited reasoner (i.e. none or transitive), some of the ancestors or descendants of a term may be missed. To compensate for that, do a few additional direct checks.
PavlidisLab · Sep 18, 2023 · f31c6c7 · f31c6c7
1 parent 6f62477
commit f31c6c7
Showing 1 changed file with 42 additions and 0 deletions.
diff --git a/src/ubic/basecode/ontology/jena/JenaUtils.java b/src/ubic/basecode/ontology/jena/JenaUtils.java
@@ -8,6 +8,8 @@
 import com.hp.hpl.jena.util.iterator.Filter;
 import com.hp.hpl.jena.util.iterator.UniqueExtendedIterator;
 import org.apache.commons.lang3.time.StopWatch;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 
 import javax.annotation.Nullable;
 import java.util.*;
@@ -18,7 +20,25 @@
 
 class JenaUtils {
 
+    protected static Logger log = LoggerFactory.getLogger( JenaUtils.class );
+
     public static Collection<OntClass> getParents( OntModel model, Collection<OntClass> ontClasses, boolean direct, @Nullable Set<Restriction> additionalRestrictions ) {
+        Collection<OntClass> parents = getParentsInternal( model, ontClasses, direct, additionalRestrictions );
+        if ( !direct && !parents.isEmpty() && additionalRestrictions != null && !supportsAdditionalPropertyInference( model ) ) {
+            // if there are some missing direct parents, revisit the hierarchy
+            Set<OntClass> parentsToRevisit = new HashSet<>( parents );
+            while ( !parentsToRevisit.isEmpty() ) {
+                log.debug( "Revisiting the direct parents of {} terms...", parentsToRevisit.size() );
+                parentsToRevisit = new HashSet<>( getParentsInternal( model, parentsToRevisit, true, additionalRestrictions ) );
+                parentsToRevisit.removeAll( parents );
+                log.debug( "Found {} missed parents.", parentsToRevisit.size() );
+                parents.addAll( parentsToRevisit );
+            }
+        }
+        return ontClasses;
+    }
+
+    private static Collection<OntClass> getParentsInternal( OntModel model, Collection<OntClass> ontClasses, boolean direct, @Nullable Set<Restriction> additionalRestrictions ) {
         ontClasses = ontClasses.stream()
                 .map( t -> t.inModel( model ) )
                 .filter( t -> t.canAs( OntClass.class ) )
@@ -65,6 +85,23 @@ public static Collection<OntClass> getParents( OntModel model, Collection<OntCla
     }
 
     public static Collection<OntClass> getChildren( OntModel model, Collection<OntClass> terms, boolean direct, @Nullable Set<Restriction> additionalRestrictions ) {
+        Collection<OntClass> children = getChildrenInternal( model, terms, direct, additionalRestrictions );
+        // collect any remaining children, this is useful if inference is incomplete (i.e. if not using a suitable OWL reasoner)
+        if ( !direct && !children.isEmpty() && additionalRestrictions != null && !supportsAdditionalPropertyInference( model ) ) {
+            // if there are some missing direct children, revisit the hierarchy
+            Set<OntClass> childrenToRevisit = new HashSet<>( children );
+            while ( !childrenToRevisit.isEmpty() ) {
+                log.debug( "Revisiting the direct parents of {} terms...", childrenToRevisit.size() );
+                childrenToRevisit = new HashSet<>( JenaUtils.getChildrenInternal( model, childrenToRevisit, true, additionalRestrictions ) );
+                childrenToRevisit.removeAll( children );
+                log.debug( "Found {} missed children.", childrenToRevisit.size() );
+                children.addAll( childrenToRevisit );
+            }
+        }
+        return children;
+    }
+
+    public static Collection<OntClass> getChildrenInternal( OntModel model, Collection<OntClass> terms, boolean direct, @Nullable Set<Restriction> additionalRestrictions ) {
         terms = terms.stream()
                 .map( t -> t.inModel( model ) )
                 .filter( t -> t.canAs( OntClass.class ) )
@@ -103,6 +140,11 @@ public static Collection<OntClass> getChildren( OntModel model, Collection<OntCl
         return result;
     }
 
+    private static boolean supportsAdditionalPropertyInference( OntModel model ) {
+        return model.getReasoner().supportsProperty( model.getProfile().SOME_VALUES_FROM() )
+                && model.getReasoner().supportsProperty( model.getProfile().ALL_VALUES_FROM() );
+    }
+
     public static Resource getRestrictionValue( Restriction r ) {
         if ( r.isSomeValuesFromRestriction() ) {
             return r.asSomeValuesFromRestriction().getSomeValuesFrom();